Perl 6 - the future is here, just unevenly distributed

IRC log for #puppet-openstack, 2014-02-27

| Channels | #puppet-openstack index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:05 ryanycol_ joined #puppet-openstack
00:48 ilbot3 joined #puppet-openstack
00:48 Topic for #puppet-openstack is now Place to collaborate on Puppet/OpenStack tools: logs at http://irclog.perlgeek.de/puppet-openstack/today
00:59 ryanycoleman joined #puppet-openstack
01:07 kitp joined #puppet-openstack
01:12 kitp joined #puppet-openstack
01:12 xarses joined #puppet-openstack
01:15 prad joined #puppet-openstack
01:15 rmoe joined #puppet-openstack
01:27 kkirkpatrick joined #puppet-openstack
01:27 lnxnut joined #puppet-openstack
01:40 mjblack joined #puppet-openstack
01:55 mkoderer joined #puppet-openstack
02:03 fandi joined #puppet-openstack
02:16 xingchao joined #puppet-openstack
02:31 thumpba joined #puppet-openstack
02:53 openstackgerrit Matt Fischer proposed a change to stackforge/puppet-keystone: LDAP: Add support for the ldap options  https://review.openstack.org/76632
02:58 thumpba joined #puppet-openstack
03:14 thumpba_ joined #puppet-openstack
03:21 sputnik13 joined #puppet-openstack
03:22 thumpba joined #puppet-openstack
03:36 markvoelker1 joined #puppet-openstack
03:52 mgagne joined #puppet-openstack
04:09 rmoe joined #puppet-openstack
04:29 lnxnut joined #puppet-openstack
04:42 britthouser joined #puppet-openstack
04:51 openstackgerrit Matt Fischer proposed a change to stackforge/puppet-keystone: LDAP: Add support for the ldap options  https://review.openstack.org/76632
05:03 michchap markvoelker: markvoelker1: ping
05:03 markvoelker1 michchap: pong
05:04 michchap markvoelker: I am struggling to think of a use for InternalURL being different to PublicURL, and similar for AdminURL except in the case of keystone.
05:05 michchap markvoelker: I ask because I don't really see the need to bind all the internal urls to the internal vip as well as the public vip, when all the control services can just use the public one.
05:06 michchap markvoelker: but if you can explain when that's actually helpful, I will refrain from re-mapping everything (eg nova) from using the public one (eg neutron api url)
05:06 michchap s/from/to
05:09 michchap markvoelker1: ^
05:09 markvoelker1 michchap: so is the question really about segreation of networks (say, exposing control plane services only to a "trusted" net)?
05:10 michchap markvoelker1: I think what I'm saying is that internalURL isn't exposing anything different to publicurl so just use public when the control plane services need it
05:11 markvoelker1 I'm more thinking of cases where the ports for each might be bound on different interfaces/networks
05:12 markvoelker1 E.g. so internalURL might be on a network with a different SLA than public
05:12 michchap they're the same port for internalurl/publicurl for everything
05:12 britthouser joined #puppet-openstack
05:12 markvoelker1 In the current datamodel you mean?
05:15 michchap I mean they're providing the same service and by default they're on the same port.
05:16 markvoelker1 yeah, and the majority of deployments I've seen do set them the same.  They can be configured differently though, no?
05:18 markvoelker1 So, I guess what I'm getting at is: IMHO I don't know that there's a real practical reason not to do it, but what's driving the desire to make the change?
05:18 markvoelker1 Just simplification of the mappings?
05:19 michchap it means I have to add  all the internalurls as loadbalanced ports on the private VIP when they are already available on the public VIP
05:20 michchap which is not a huge issue, it just seems a bit pointless
05:20 michchap I don't know how to configure them differently - maybe via policy
05:21 michchap like this: https://github.com/CiscoSystems/puppet_openstack_builder/blob/master/data/data_mappings/common.yaml#L161
05:21 michchap I can't see any advantage to that being internalurl over publicurl. I want to have a controller internal address for services I don't want to make public, but the API endpoint are already going to be public.
05:22 michchap the next line down makes sense because it's the admin auth url, which should be on the admin endpoint
05:23 michchap s/line/mapping
05:24 markvoelker1 Got it.  Yeah, I guess I can't really think of any actual deployments where it would break anything.
05:24 michchap so if I change those mappings to use the public address instead of the private, I don't think I will break any deployments, because I don't know of anyone acutally using the internalurl for anything - it's the same uri as the publicurl.
05:25 michchap I'll double check compressed HA isn't affected - it's the only one that might be.
05:26 michchap that whole block is actually a bit awkward - if we split out the controller to separate nodes, it falls apart.
05:28 markvoelker1 yep, a bit
05:31 michchap cool. Will have a patch for this soon and you can check it doesn't break stuff.
05:31 michchap actually daneyon is setting up listens for the internal ports
05:32 michchap I'll just do that. I don't want to make compressed HA broken while you're trying to use it.
05:32 markvoelker1 Ok.  We can always optimize later.
05:34 michchap We should try to clean all this crap for your icehouse release
05:34 michchap get allt he terminology consistent
05:34 michchap public/private/mgmt/internal
05:34 michchap all that stuff
05:35 markvoelker1 Yeah, upstream makes it confusing enough without us adding more =p
05:35 michchap I've used 4 networks for my scenario: mgmt, public, internal and data
05:35 michchap yep
05:35 michchap take a read of the RDO docs and the Rackspace deploy guide and stuff
05:36 michchap I like to imagine writing that kind of doc as I'm working, otherwise I end up using different terms
05:36 michchap and node vs server is all over the place too
05:39 markvoelker1 michchap: on that note, we were actually tossing around the idea of adding a directory to POB with sets of sample configs w/READMEs
05:39 markvoelker1 (and diagrams, lest the SE's come after me with torches and pitchforks)
05:40 michchap chuck it under contrib :)
05:40 michchap I'm all for it
05:40 michchap btw vagrant is adding a share capability
05:41 openstackgerrit A change was merged to stackforge/puppet_openstack_builder: Remove git remote references  https://review.openstack.org/76337
05:41 michchap which would be a neat way for me to demo these changes to you without having to sort out all the module changes.
05:42 markvoelker1 michchap: there's an argument to made that the OOB settings should "just work" and therefore sample configs are unnecessary...
05:42 michchap because everyone has the same SSL cert amirite?
05:42 markvoelker1 ..but, reality is it's impsosible to guess what NICs people might or might not have in the real world (as a for instance)
05:42 markvoelker1 exactly
05:43 markvoelker1 One could maybe hope to get most of the way there for AIO, but that's about it
05:43 michchap my goal is to make the next release a bit more understandable by breaking up some up the mappings to smaller things. Like the distinction between bind addresses and controller addresses
05:44 michchap but it's never going to work ootb unless you're building in a VM and can control the network config
05:44 michchap there's also the question of whether certain things qualify as working, like using file backed glance store.
05:45 britthouser joined #puppet-openstack
05:48 markvoelker1 agreed
05:48 markvoelker1 If you're interested in hacking up the mappings a bit, I may have some useful 3rd part input in a while
05:49 markvoelker1 Some of the folks from WebEx/etc have been looking at it and commented to me that they found it a bit confusing and may have some ideas
05:50 markvoelker1 They're all buried under some deadlines of their own for a bit, but might be some good end-user feedback for Icehouse.
05:56 xingchao joined #puppet-openstack
06:07 openstackgerrit Mark T. Voelker proposed a change to stackforge/puppet_openstack_builder: Enable Neutron LBaaS and FWaaS services  https://review.openstack.org/76764
06:18 toabctl joined #puppet-openstack
06:33 dgollub joined #puppet-openstack
06:39 markvoelker1 joined #puppet-openstack
06:48 saju_m joined #puppet-openstack
07:03 britthouser joined #puppet-openstack
07:12 dgollub joined #puppet-openstack
07:13 sputnik13 joined #puppet-openstack
07:15 thumpba joined #puppet-openstack
07:30 britthouser joined #puppet-openstack
07:43 dgollub joined #puppet-openstack
08:10 saju_m joined #puppet-openstack
08:11 britthouser joined #puppet-openstack
08:17 bkero joined #puppet-openstack
08:18 sputnik13 joined #puppet-openstack
08:22 kitp joined #puppet-openstack
08:25 britthouser joined #puppet-openstack
08:26 saju_m joined #puppet-openstack
08:30 bauzas joined #puppet-openstack
08:38 markvoelker1 joined #puppet-openstack
08:43 dgollub joined #puppet-openstack
08:44 britthouser joined #puppet-openstack
09:19 mmagr joined #puppet-openstack
09:23 kitp joined #puppet-openstack
09:25 britthouser joined #puppet-openstack
10:05 britthouser joined #puppet-openstack
10:24 kitp joined #puppet-openstack
10:34 otherwiseguy joined #puppet-openstack
10:35 saju_m joined #puppet-openstack
10:46 britthouser joined #puppet-openstack
11:17 thumpba joined #puppet-openstack
11:17 britthouser joined #puppet-openstack
11:25 kitp joined #puppet-openstack
11:50 britthouser joined #puppet-openstack
11:57 thumpba joined #puppet-openstack
12:07 britthouser joined #puppet-openstack
12:25 kitp joined #puppet-openstack
12:28 dmsimard joined #puppet-openstack
12:30 openstackgerrit Emilien Macchi proposed a change to stackforge/puppet-nova: compute: RBD backend support  https://review.openstack.org/76462
12:30 EmilienM markvoelker: i've modified to support Havana ^
12:30 EmilienM i don't know if i'm allowed to modify a backport
12:30 EmilienM if not, please -2 and I'll do another patchset
13:09 morazi joined #puppet-openstack
13:13 openstackgerrit Emilien Macchi proposed a change to stackforge/puppet-neutron: Debian: fix package and service names  https://review.openstack.org/76858
13:13 EmilienM markvoelker: ^
13:26 kitp joined #puppet-openstack
13:32 finnx joined #puppet-openstack
13:37 mjblack joined #puppet-openstack
13:37 britthouser joined #puppet-openstack
13:41 britthou_ joined #puppet-openstack
13:45 michchap markvoelker: lol I just realised network_controller doesn't include neutron::server
13:46 EmilienM michchap: indeed :P
13:51 markvoelker michchap: wow.  That's pretty awesome. =p
13:51 dprince joined #puppet-openstack
13:51 markvoelker I probably would have noticed if I hadn't submitted that last patch at 1am on a day when I had a 4am meeting =p
13:51 markvoelker EmilienM: sure, I'll have a look
13:54 EmilienM markvoelker: yeah, i would like to have a look at both issues we have (compute/rbd and neutron/package/service names for Debian)
13:54 EmilienM markvoelker: would like you*
13:56 markvoelker EmilienM: I have no  problem with the rbd backport one.
13:56 EmilienM markvoelker: I changed the backport
13:56 EmilienM I don't know if I'm allow to do that
13:57 britthouser joined #puppet-openstack
13:57 markvoelker IMHO it's better to do it this way (as a cherry pick plus minor modification) than to do it as a whole separate commit.  Less chance of manual changes causing issues, and points reviewers to the original as a base.
13:58 markvoelker I don't know if there's any real consensus on the matter, but if so I'm sure the next core dev to review it will pipe up about it. =)
14:01 EmilienM markvoelker: far enough, let's dot it like it's currently
14:02 kitp joined #puppet-openstack
14:04 britthouser joined #puppet-openstack
14:05 fandi joined #puppet-openstack
14:10 michchap EmilienM: what is normally in /etc/neutron/plugin.ini?
14:10 EmilienM michchap: it's a symlink of the plugin you actually use
14:11 EmilienM ex: ml2_config.ini
14:11 michchap hmm should that be made by the package?
14:11 openstackgerrit Chris Ricker proposed a change to stackforge/puppet-neutron: neutron-metering-agent service name is wrong  https://review.openstack.org/76661
14:12 EmilienM michchap: in RHEL, it is
14:12 EmilienM not for ubuntu/debian
14:12 michchap Ah I see it in our manifests...I wonder why I'm not seeing it.
14:12 michchap I just made a clean box and it's not there. odd.
14:14 openstackgerrit Chris Ricker proposed a change to stackforge/puppet-neutron: VPNaaS service name is wrong  https://review.openstack.org/76662
14:19 openstackgerrit Chris Ricker proposed a change to stackforge/puppet-neutron: VPNaaS service name is wrong  https://review.openstack.org/76662
14:23 michchap markvoelker: you can tell we never used the network_controller group: it's missing this too: "%{network_service}::plugins::%{network_plugin}"
14:32 britthouser joined #puppet-openstack
14:37 michchap markvoelker: Are you able to add me as a reviewer now?
14:37 openstackgerrit A change was merged to stackforge/puppet-neutron: VPNaaS service name is wrong  https://review.openstack.org/76662
14:37 markvoelker michchap: yes, just noticed that this morning
14:38 michchap markvoelker: looks like fungi did some magic. No errors now?
14:38 markvoelker michchap: no errors
14:40 * markvoelker runs off to a meeting
14:49 92AAAA5TT joined #puppet-openstack
14:57 britthouser joined #puppet-openstack
15:01 openstackgerrit Michael Chapman proposed a change to stackforge/puppet-cinder: Add support for puppetlabs-mysql 2.2  https://review.openstack.org/74048
15:04 openstackgerrit Michael Chapman proposed a change to stackforge/puppet-heat: Add support for puppetlabs-mysql 2.2  https://review.openstack.org/74224
15:06 mjblack joined #puppet-openstack
15:08 mjblack joined #puppet-openstack
15:09 openstackgerrit Michael Chapman proposed a change to stackforge/puppet-keystone: Add support for puppetlabs-mysql 2.2  https://review.openstack.org/76501
15:13 dmsimard1 joined #puppet-openstack
15:14 lnxnut joined #puppet-openstack
15:16 dprince_ joined #puppet-openstack
15:28 ianw joined #puppet-openstack
15:28 mmagr joined #puppet-openstack
15:30 openstackgerrit Michael Chapman proposed a change to stackforge/puppet-glance: Add support for puppetlabs-mysql 2.2  https://review.openstack.org/74061
15:33 openstackgerrit Michael Chapman proposed a change to stackforge/puppet-neutron: Add support for puppetlabs-mysql 2.2  https://review.openstack.org/74227
15:38 fandi joined #puppet-openstack
15:38 openstackgerrit A change was merged to stackforge/puppet-neutron: neutron-metering-agent service name is wrong  https://review.openstack.org/76661
15:39 kkirkpatrick joined #puppet-openstack
15:42 openstackgerrit Michael Chapman proposed a change to stackforge/puppet-nova: Add support for puppetlabs-mysql 2.2  https://review.openstack.org/74223
15:49 britthouser joined #puppet-openstack
15:51 openstackgerrit Michael Chapman proposed a change to stackforge/puppet-nova: Add support for latest puppetlabs-rabbitmq  https://review.openstack.org/76896
15:54 openstackgerrit Michael Chapman proposed a change to stackforge/puppet-nova: Add nova-objectstore bind address  https://review.openstack.org/76897
15:54 prad_ joined #puppet-openstack
15:58 openstackgerrit Michael Chapman proposed a change to stackforge/puppet-nova: Add support for puppetlabs-mysql 2.2  https://review.openstack.org/74223
15:59 thumpba joined #puppet-openstack
16:07 thumpba joined #puppet-openstack
16:21 britthouser joined #puppet-openstack
16:30 dgollub joined #puppet-openstack
16:47 sputnik13 joined #puppet-openstack
16:49 britthouser joined #puppet-openstack
16:55 openstackgerrit Mark T. Voelker proposed a change to stackforge/puppet_openstack_builder: Expose controller_public_protocol  https://review.openstack.org/76914
17:00 morazi joined #puppet-openstack
17:03 mgagne joined #puppet-openstack
17:04 openstackgerrit Donald Talton proposed a change to stackforge/puppet-nova: Added the option to create the nova user before installation  https://review.openstack.org/76298
17:04 mgagne1 joined #puppet-openstack
17:06 openstackgerrit Matt Fischer proposed a change to stackforge/puppet-keystone: LDAP: Add support for the ldap options  https://review.openstack.org/76632
17:09 britthouser joined #puppet-openstack
17:10 thumpba joined #puppet-openstack
17:26 bauzas joined #puppet-openstack
17:27 britthouser joined #puppet-openstack
17:27 rmoe joined #puppet-openstack
17:29 fandi joined #puppet-openstack
17:31 openstackgerrit Donald Talton proposed a change to stackforge/puppet-nova: Added the ability to manage the nova user  https://review.openstack.org/76298
17:33 Aimon_ joined #puppet-openstack
17:34 britthouser joined #puppet-openstack
17:45 dtalton joined #puppet-openstack
17:50 xarses joined #puppet-openstack
17:51 britthouser joined #puppet-openstack
17:52 xarses joined #puppet-openstack
18:02 dtalton fvollero, i have a rhel package question whenever you have a moment
18:06 fvollero dtalton: sure mate, shoot
18:07 dtalton thanks. my question is that if i already create a user, eg nova, before installing any nova packages, will installing the package 1. use that existing user and 2. modify that users uid/gid to the right "redhat" one?
18:11 angdraug joined #puppet-openstack
18:34 britthouser joined #puppet-openstack
18:37 saju_m joined #puppet-openstack
18:41 csschwe joined #puppet-openstack
18:49 fvollero sigh
18:49 fvollero dalton logged out
18:51 EmilienM fvollero: he could read backlog on http://irclog.perlgeek.de/puppet-openstack/today
18:51 saju_m joined #puppet-openstack
18:52 markvoelker fvollero: pretty sure his question is rooted in this:  https://review.openstack.org/#/c/76298/
18:52 EmilienM markvoelker: nice catch !
18:52 fvollero EmilienM: Coolio :)
18:52 csschwe joined #puppet-openstack
18:52 xarses markvoelker: fvollero id guess it is too
18:53 markvoelker fvollero: IIRC, uid/gid aren't preassigned in the Ubuntu packaging, but are (?) in Red Hat packaging.  So question is how a change like this would affect Red Hat platforms.
18:53 fvollero dtalton: the answer is: it will change the user as it is in the spec file
18:55 fvollero markvoelker: I should investigate  more, but I am sure in the spec file that the rpm read/parse he will run a check if the user exist, but i am worried that if have a different id the owner will be the numeric id more than the user.
18:55 markvoelker fvollero: yeah, that definitely has some implications on this patch
18:55 xarses fvollero: it depends on the rpm package, proper packaging looks to see if the user exists first, and if it does, doesn't create one, if it doesn't it tries to create one with it's default uid/gid and if those aren't available, it lets the system pick the next
18:56 dtalton joined #puppet-openstack
18:56 EmilienM fvollero: back ^
18:57 dtalton2 joined #puppet-openstack
18:58 dtalton2 ok so it sounds as if I need to inspect the nova-common package, and probably do an os check in the manifest too
18:59 fvollero dtalton2: there're on github the specfile of all the openstack components
19:00 fvollero xarses: it's pixelb that do most of the packaging and he's very very careful on package things in the proper way :)
19:00 xarses dtalton2: I still don't like that we are removing the nova user in the case that we aren't managing the uid/gid. See my review on patch set 2
19:00 xarses removing in that, we aren't defining it any longer
19:01 dtalton2 on debian, it's create by the nova-common package if it doesn't already exist
19:01 dtalton2 and if it does, it's left alone
19:01 xarses It's a puppet issue, not "is it created" issue
19:01 xarses I cant use "User['nova']" now in my manifests because it wont be defined if i dont set the uid
19:04 dtalton how about i stick an else in there and put the old user/group nova setup there?
19:05 xarses see my comment, its easy to do, you just place your chains in the if not in user and group
19:05 xarses well actually the group one is fine
19:06 xarses you can leave the uid / gid stuff you have, puppet wont change the value if undef is passed
19:09 britthouser joined #puppet-openstack
19:15 sputnik13 joined #puppet-openstack
19:39 kitp joined #puppet-openstack
19:39 britthouser joined #puppet-openstack
19:41 mgagne joined #puppet-openstack
19:56 openstackgerrit Donald Talton proposed a change to stackforge/puppet-nova: Added the ability to manage the nova uid/gid.  https://review.openstack.org/76298
19:57 openstackgerrit Donald Talton proposed a change to stackforge/puppet-nova: Added the ability to manage the nova uid/gid.  https://review.openstack.org/76298
20:02 Aimon joined #puppet-openstack
20:08 britthouser joined #puppet-openstack
20:46 xarses_ joined #puppet-openstack
20:51 ianw joined #puppet-openstack
21:10 britthouser joined #puppet-openstack
21:23 ryanycol_ joined #puppet-openstack
21:34 britthouser joined #puppet-openstack
21:53 lnxnut joined #puppet-openstack
21:54 lnxnut joined #puppet-openstack
21:58 lnxnut_ joined #puppet-openstack
21:59 britthouser joined #puppet-openstack
22:00 ryanycoleman joined #puppet-openstack
22:01 iheartwhiskey joined #puppet-openstack
22:16 ryanycoleman joined #puppet-openstack
22:26 britthouser joined #puppet-openstack
22:27 ryanycoleman joined #puppet-openstack
22:29 lnxnut joined #puppet-openstack
22:40 bauzas joined #puppet-openstack
22:41 britthou_ joined #puppet-openstack
22:44 lnxnut joined #puppet-openstack
23:11 xarses joined #puppet-openstack
23:12 mgagne joined #puppet-openstack
23:33 openstackgerrit Michael Chapman proposed a change to stackforge/puppet-nova: Add support for puppetlabs-mysql 2.2  https://review.openstack.org/74223
23:45 xarses joined #puppet-openstack
23:45 xarses joined #puppet-openstack
23:45 mattymo joined #puppet-openstack

| Channels | #puppet-openstack index | Today | | Search | Google Search | Plain-Text | summary